Envisioning knowledge: tightly coupling knowledge analysis and visualization

Analysis and visualization techniques share a symbiotic relationship when it comes to making sense of datasets. Particularly for large datasets, the coupling of data analysis and data visualization is often beneficial. While a gamut of data analysis and visualization techniques exist, it is often problematic to identify what combination of techniques is good for what task to provide maximum insight into a dataset. This dissertation introduces, exemplifies and validates a Data Analysis and Visualization Taxonomy, called 'DA-Vis taxonomy', that provides guidance to the selection of complementary analysis and visualization techniques. The DA-Vis taxonomy is validated by demonstrating its utility to develop new visualizations for real world applications. Further, the new taxonomy is applied to systematically describe and classify couplings of data analysis and visualization techniques in prior work. A user-study that evaluates the usability of the DA-Vis taxonomy was also conducted and is reported here. The intellectual contributions of this dissertation include a flexible DA-Vis layout schema that can be used to tightly couple complementary data analysis and visualization techniques. This dissertation also shows a visionary computational diagnostic tool developed for data analysis and visualization of clinical data. Techniques used to generate meaningful knowledge management visualizations from a dataset are presented as a part of this dissertation. This dissertation concludes with a discussion of the broader impacts of the DA-Vis taxonomy, the computational diagnostic tool, and knowledge management maps.